The sustainability probability for the critical Derrida-Retaux model (Q2073186)

Consider an integer \(m \geq 2\), and let \(X_{0} \geq 0\) be a random variable taking values in \(\mathbb{Z}_{+}:=\{0,1,2, \ldots\}\) and satisfying \(\mathbf{E}\left(X_{0} m^{X_{0}}\right)<\infty\). Define the following recurrence relation: for all \(n \geq 0\), \[ X_{n+1}=\left(X_{n}^{(1)}+\cdots+X_{n}^{(m)}-1\right)^{+}, \] where \(X_{n}^{(i)}, i \geq 1\), are independent copies of \(X_{n}\) and \( x^{+}:=\max \{x, 0\}\) for all \( x \in \mathbb{R}\). The authors define this model, and their basic question is how the distribution of \(X_{n}\) behaves when \(n\) is sufficiently large. The critical case is \[(m-1) \mathbf{E}\left(X_{0} m^{X_{0}}\right)=\mathbf{E}\left(m^{X_{0}}\right)<\infty.\] The main result of the paper is that at criticality, the probability \(\mathbf{P}\left(X_{n}>0\right)\) behaves like \(n^{-2+o(1)}\) as \(n\) goes to infinity.